The problem of improving the efficiency and accuracy of measurements and regulation of liquid flow rates attracts high attention in various branches of the contemporary economy [1,2,3]. It is also of great importance to be able to perform high-precision dosing of small volumes of liquids for the purposes of chemical and biochemical industry and medicine, e.g., for the regulation of consumption of the drug solutions in infusion systems [4], liquid chemical substances, reagents, catalysts, etc. [5].

For the automatic high-precision regulation of the consumption of liquids with different physicochemical properties, it is necessary to have precision measuring instruments realizing the noncontact procedure of measurements in the automatic mode, which is especially important for dosing chemically active, toxic, and poisonous liquids.

Materials and methods. For the solution of the problem of precision measurements and regulation of low liquid flow rates, we have developed a device whose scheme and detailed description can be found in [6]. The device contains a gauge of low liquid flow rates [7, 8] connected with an analog-to-digital converter (ADC), a microcontroller, an input block, a display, and a flow-rate regulator mounted on the gauge. The gauge of low liquid flow rates [6] consists of the case and the following elements mounted inside it: fastening rings with a source of radiation and a photodetector, holders with a dispersing lens made in the form of a glass cylinder and a converging lens, and a plate with horizontal slot. A translucent element with inlet and outlet branch pipes for the liquid and a drop generator are placed in the hole of the case perpendicularly to its axis. Two lenses and two plates form an optical system with the horizontal slot. This system converts the light flux of the source of radiation and forms a measuring plane inside the translucent element. A regulator of liquid flow rates is mounted on the outlet branch pipe of the gauge of low liquid flow rates. The regulator contains a servodrive (mounted on a supporting arm fastened to the outlet branch pipe of the transducer) and a lever.

The device aimed at the automatic control of low liquid flow rates operates in the following way.

With the help of the input device, the required value of the liquid flow rate, which does not exceed the maximum admissible level, is introduced in the memory of the microcontroller. These data are represented in the display whose input is connected with the output of the microcontroller. Then we attach tubes to the inlet and outlet branch pipes of translucent cylinder. The liquid from the inlet branch pipe in the form of drops comes from the drop generator, passes through the measuring plane of translucent element, and is removed through the outlet branch pipe. The radiation coming from the source passes through the dispersing lens, changes its direction to almost parallel and is converted into the light flux. Then the light flux comes to the horizontal slot made in the plate, where it is formed as a measuring plane passing through the translucent element with partial refraction and scattering on the walls. When a drop moves through the measuring plane, the light flux is distorted and its intensity decreases. After the translucent element, the light flux passes through the horizontal slot in the plate and the converging lens. This enables us to focus the light flux in the form of a beam, which guarantees the reliable hitting of the photosensitive element of photodetector by this beam. In this case, the signal from source of radiation comes to the photodetector through the optical system in the form of a distorted light beam and is transmitted to ADC, where it is processed and delivered to the input of the microcontroller. Depending on the frequency of coming signals, the microcontroller determines the current liquid flow rate, compares it with the specified value, and displays information in the screen.

If the current liquid flow rate is higher than the specified value, then a signal from the output of the microcontroller is transmitted to the regulator whose servodrive acts through the lever upon the outlet branch pipe of the gauge of low liquid flow rates and partially blocks the branch pipe (by decreasing its opening). As a result, the liquid flow rate decreases. If the current liquid flow rate is lower than the specified value, then the signal comes in the same way to the opening of the outlet branch pipe (and increases the opening). As a result, the liquid flow rate also increases.

For the investigation of the technical characteristics and accuracy of developed device and evaluation of the degree of influence of various parameters of liquid on the process of measurements and regulation of the flow rate, we designed and produced an experimental installation whose schematic diagram is shown in Fig. 1. The installation contains several reservoirs 8 with samples of investigated liquids to which one can successively supply every liquid (without mixing with other liquids) with the help of mechanical flow-rate regulators 7 and mixer 9. Our installation also contains the gauges of small flow rates 6 and transmittance of liquids 10, a control unit 5 based on a microcontroller with an AT91SAM3X8E Atmel microprocessor, including an LCD-1602 symbol liquid-crystal display 4 for reproducing the data and a TSOP-1736 infrared sensor 2. An MG996r servomechanism 11 is used for the precise automatic regulation of the flow rate of a given liquid or complete blocking of the outlet branch pipe at the end of a cycle of portion dosing. The required flow rate is specified with the help of the control panel 13 and regulated by the microcontroller. A high-precision Diamond balance 1 equipped with a reservoir for liquids 12 is used for the comparison of the obtained experimental results (based on the program analysis of the data of scanning each drop) with the actual values of the flow rate with regard for the density of liquids. All information obtained in the course of experiments is transmitted through a virtual COM port to an ASUS-U38N notebook 3 with Microsoft Windows 10 software and processed by using the Microsoft Excel software.

Fig. 1.
figure 1

Experimental installation (schematic diagram): 1) high-precision balance; 2) infrared sensor; 3) notebook; 4) symbol liquid-crystal display; 5) control unit; 6) gauge of low liquid flow rates; 7) mechanical flow-rate regulators; 8) reservoirs with samples of liquids; 9) mixer; 10) gauge of liquid transmittance; 11) servomechanism; 12) reservoir for liquids; 13) control panel.

Full size image

We carried out a series of experiments with samples of liquids with different dynamic viscosities, densities, and refraction coefficients: 9% physiological solution (with a density of 1.005 g/cm3), 9% physiological solution with a pigment (with a density of 1.005 g/cm3), glycerol (with a density of 1.261 g/cm3), a glycerol–alcohol solution in a ratio 3:2 (with a density of 1.075 g/cm3), and a glycerol–alcohol solution in a ratio 3:2 containing a pigment (with a density of 1.075 g/cm3). The experiments were performed under standard conditions (at a temperature of 25°C). In this case, for liquids not susceptible to rapid evaporation (water and physiological solution), the process of their evaporation was neglected due to its weak influence (much lower than the measurement error) on the experimental results. For rapidly evaporating liquids (alcohols, etc.), the results of measurements were recorded after a certain period of time from the onset of the experiment, which enabled us to stabilize the level of humidity in the zone of translucent element and in the closed reservoir for the collection of liquids (in the case of saturation with vapors) and to minimize the influence of evaporation of liquids at constant temperature in short-term experiments.

The physiological solutions (pure and with pigments) have almost identical densities, dynamic viscosities, and other physical characteristics, except the transmittance (due to different refraction coefficients). The glycerol–alcohol solutions with and without pigments also differ solely by their transmittance.

In a single series of experiments, we studied 100 drops of the described liquids. With the help of the developed specialized software, every drop was scanned by the light flux from the source of radiation 200 times in the course of its flight through the measuring plane of the gauge of low liquid flow rates. After this, we determined the mean value (in mV) of the level of output signal of the ADC for each of 200 points of scanning of the drop, and the corresponding plots were constructed. After a series of tests, we weighed the liquid (the step of balance discretization was 0.001 g; the balance was calibrated with the help of a laboratory weight within the working measurement range).

The current volume flow rate of liquid was determined by analyzing the integral component obtained according to the difference of the output signals of ADC in scanning of the measuring cavity by the light flux in the course of the flight of a drop of the analyzed liquid and in the absence of drops. Here, we have introduced a special coefficient taking into account the transmittance of liquid and empirically determined for different liquids after preliminary tests with the use of high-precision balance (by means of calibration). Our experimental installation also contained an additional gauge of low liquid flow rates [7] mounted on the pipe higher than the drop generator and making it possible to preliminarily determine the currently analyzed liquid.

As the translucent element in our experimental installation, we used a standard device for the injection of blood substitutes and infusion solutions (dropper) [9]. To regulate the frequency of drop generation, we changed the opening of the outlet branch pipe. As a result, the pressure in the top part of the translucent element not filled with liquid varied, and, as a result, the process of drop generation was decelerated or accelerated. The influence of changes in pressure in the top part of the translucent element on the parameters of generated drops was much lower than the measurement error and, hence, was neglected in our tests.

For the preliminary investigation of the accuracy of liquid dosing at different flow rates (specified by the position of the lever of servomechanism), we selected the required dose in drops (100) for different liquids in the display. After dosing of the required amount, the delivery of liquid was blocked by the servomechanism and the total number of drops in the dose was recorded and displayed in the screen. The obtained dose of liquid was weighed and its mass was compared with the computed theoretical value (with regard for the average mass of a single drop of the analyzed liquid).

Discussion of the results. As a result of weighing, we determined the mean mass of a single drop of each liquid: 0.050 g for a 9% physiological solution (the volume of a drop was 0.050 cm3), 0.022 g for a glycerol–alcohol mixture (0.020 cm3), and 0.028 g for glycerol (0.022 cm3). A drop of glycerol–alcohol mixture has the minimal volume.

We can observe the variations of the intensity of light flux coming to the light-sensitive element caused by the flights of the drops of different liquids across the measuring plane of the gauge of low flow rates in the form of variations of the curves U(t) of the output signal of the 12-digit ADC (Fig. 2). Due to different levels of surface tension of the analyzed liquids (72.75 mN/m for the physiological solution and 59.4 mN/m for glycerol), the rate of separation of glycerol drops from the drop generator is higher and, hence, the volume of the drops is lower than for the drops of physiological solution. Moreover, the glycerol drops have a much higher dynamic viscosity as compared with the physiological solution (1495 mPa·sec and 1 mPa·sec at 20°C) and, hence, take the shape of a ball after separation from the drop generator slower than the drops of physiological solution, which is demonstrated by the lower steepness of curve 1 in Fig. 2.

Fig. 2.
figure 2

Mean values of the output signals of ADC caused by the flights of liquid drops: 1–3) pure glycerol, glycerol–alcohol solution, and glycerol–alcohol solution with pigment, respectively; 4, 5) physiological solutions (pure and with pigment, respectively).

Full size image

We determined the area of the figure bounded below by the curve U(t) for a given liquid and bounded above by the level U = 2019 mV corresponding to the mean output voltage of the ADC in the absence of drops in the scanning zone. The reduced (with regard for the refraction coefficient of liquid) value of this area enables one to make indirect conclusions concerning the main properties of the analyzed liquid (density and dynamic viscosity) and to determine its type in the presence of the database of liquids.

The preliminary investigation of the accuracy of dosing for different flow rates of liquids shows that, in all our tests, the number of drops collected in the reservoir for liquids was equal to their specified number displayed in the screen. To find the error of the developed device for the automatic control of low liquid flow rates, we carried out several series of tests with a physiological solution. For the processing of the experimental results, we used the Kornfeld method. The width of the confidence interval for the integral values of the results of optical measurements of the mass of liquid is 1.8%. This fact enables us to judge the correctness of operation of our software and the entire device and forms a prerequisite for the creation of specialized software aimed at the automatic high-precision regulation of low liquid flow rates with a possibility of correction of the specified value of flow rate in real time. Thus, the developed device [7, 8] differs from the existing devices by the use of microcontroller receiving and processing the signals from the ADC on the current liquid flow rate and from the input block on the required flow rate and a specialized executive routine aimed at adjusting the flow rate in real time. Here, we take into account the current measurement error and minimize it within a short time interval after revealing the error.

According to the results of processing of the data on the required and current actual liquid flow rates, the microcontroller forms a control action upon the flow regulator. The regulator and gauge of low liquid flow rates adjust the current flow rate to the required value. The fact that the cycle is closed guarantees the possibility of permanent control over the liquid flow rate, its regulation, and maintaining on the required level. Thus, we manage to realize the precise automatic regulation of low liquid flow rates.

Conclusions. As the advantages of proposed device for the automatic control of low liquid flow rates, we can mention the noncontact optical procedure of measurements according to which the measured liquid does not interact with the installation and the ambient medium, which is required in the investigation and dosing of toxic, explosive, and poisonous liquids. In using this device, the minimal measured flow rates of aqueous solutions can be as low as 0.05·10–3 liter/h. It is also possible to measure the flow rate of a specific liquid without separation from the total flow and (or) deceleration of its motion, which is impossible to realize by the methods of final measurement (with application of high-precision balance, volumetric flowmeters, etc.). Thus, the application of the developed device guarantees high accuracy and reliability of the noncontact method of automatic measurements, dosing, and regulation of low flow rates for almost any liquids and creates conditions for the total automation of the process of high-precision dosing of different liquids with a possibility of control over the dosing process in real time.